2026-08-16 21:30 UTC (44131m ago) · next ~13:15 UTC

Frontier LLM API Status

Is your AI provider down right now? Multi-source, confidence-weighted status for Claude, GPT, Gemini, Grok, Groq, and more.

Maintenance0/10 operational

Our pipeline status →

↗

API

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

CHAT

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

CODE

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

Top models

  • 1Claude Opus 5Operational · stale
  • 2Claude Fable 5Operational · stale
  • 3Claude Opus 4.8Operational · stale
  • 4Claude Sonnet 5Operational · stale
  • 5Claude Opus 4.7Operational · stale

Status reflects the Anthropic API surface — not per-model telemetry. Per-model availability is derived from the OpenRouter routing edge.

API latency — 24h

Loading…
↗

API

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

REALTIME

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

Top models

  • 1GPT-5.6 SolDegraded · stale95% uptime · 24h via OpenRouter
  • 2GPT-5.6 TerraDegraded · stale83% uptime · 24h via OpenRouter
  • 3GPT-5.5Operational · stale
  • 4GPT-5.4Operational · stale
  • 5GPT-5.6 LunaDegraded · stale88% uptime · 24h via OpenRouter

Status reflects the OpenAI API surface — not per-model telemetry. Per-model availability is derived from the OpenRouter routing edge.

API latency — 24h

Loading…
↗

API

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

Top models

  • 1Gemini 3.7 FlashOperational · stale
  • 2Gemini 3.5 FlashOperational · stale
  • 3Gemini 3.6 FlashOperational · stale
  • 4Gemini 3.1 Pro PreviewOperational · stale
  • 5Gemini 3.5 Flash LiteOperational · stale

Status reflects the Google Gemini API surface — not per-model telemetry. Per-model availability is derived from the OpenRouter routing edge.

API latency — 24h

Loading…
↗

API

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

Top models

  • 1Grok 4.6Operational · stale
  • 2Grok 4.5Operational · stale
  • 3Grok Build 0.1Operational · stale
  • 4Grok 4.3Operational · stale
  • 5Grok 4.20 Multi-AgentDegraded · stale78% uptime · 24h via OpenRouter

Status reflects the xAI API surface — not per-model telemetry. Per-model availability is derived from the OpenRouter routing edge.

API latency — 24h

Loading…
↗

API

100
OperationalHigh confidence Stale

Last known value — our pipeline hasn’t refreshed this recently.

API latency — 24h

Loading…
↗

API

—
UnknownLow confidence Stale

No fresh, reliable signal right now — we won't guess a number.

API latency — 24h

Loading…
↗

API

—
MaintenanceHigh confidence Stale

Under maintenance — a planned state, not an outage.

API latency — 24h

Loading…

About ClankerStatus

ClankerStatus is an independent, multi-source status monitor for frontier LLM APIs. Instead of trusting a single provider’s status page, it aggregates official feeds, synthetic probes, and community reports into a single confidence-weighted health score — and surfaces honest uncertainty when sources disagree or data goes stale. How it works →

What is ClankerStatus?
ClankerStatus is a multi-source, confidence-weighted status dashboard for frontier LLM APIs — Claude, GPT-4o, Gemini, Grok, Groq, Fireworks AI, and Together AI. It aggregates each provider's official status page, synthetic availability probes, and trust-weighted community reports into a single health score (0–100) updated every 15 minutes.
Which AI APIs does ClankerStatus monitor?
ClankerStatus monitors seven frontier LLM providers: Anthropic Claude (status.claude.com), OpenAI (status.openai.com — ChatGPT and API), Google Gemini, xAI Grok (status.x.ai), Groq (groqstatus.com), Fireworks AI, and Together AI. Each has a dedicated status page with component-level detail and per-model availability overlays.
What does confidence-weighted AI status mean?
Each status observation is weighted by how recent it is and how much sources agree. When monitoring sources disagree by two or more severity levels, ClankerStatus surfaces 'Conflicting Signals' rather than averaging them into a misleadingly confident number. When all data expires, it shows 'Unknown' rather than a stale green. Accuracy — including honest uncertainty — is the product.
How often does AI API status update?
Status data refreshes every 15 minutes for all users, and any signed-in user can set up email alerts for status changes. The dashboard shows a live indicator when connected to real-time updates. Pro subscribers additionally receive near-real-time updates via a dedicated 1-minute monitoring queue and a personal watchlist.
What makes ClankerStatus different from provider status pages?
Provider status pages have an incentive to show green. ClankerStatus aggregates each provider's official feed, synthetic probes, and community reports, then surfaces 'Conflicting Signals' when sources disagree and 'Unknown' when data goes stale. It is the only frontier LLM status aggregator with explicit confidence scoring and uncertainty modeling — a status product that goes confidently wrong is worse than none.

Dedicated status pages: Claude status · OpenAI status · Gemini status · Grok status · Groq status · Fireworks status · Together AI status · methodology